A digital camera system is normally based on a lens system comprising a number of optical elements that image a scene onto an array of optoelectronic detector elements. As digital camera systems have evolved, these optical elements and detector arrays have been becoming progressively smaller. Unfortunately, angular resolution and number of resolvable object points typically scale with the size of an imaging system. As a result, the optical performance of such camera systems begins to suffer as the optical elements and detector elements continue to shrink.
Typically, it is desirable for the lens system to (1) collect as much of the light signal as possible over as large an aperture as possible; and (2) process the collected light signal to either form an optical image on the detector array or to encode the light signal for digital image estimation. Each detector in the detector array receives light from the lens system and converts it into an electrical signal whose magnitude is a function of light intensity. These electrical signals are then processed to develop a composite digital image of the scene and/or estimate one or more properties of the scene.
Lens system design begins by specifying targets for major performance metrics, such as angular resolution, field-of-view, depth of field, spectral range, sensitivity, dynamic range, system mass and volume. Angular resolution is generally the most significant initial metric. The best angular resolution of a lens is given by λ/A, where λ is the operating wavelength and A is the collection aperture diameter. Once the collection aperture size has been determined by this relationship, a lens is designed to achieve the remaining performance metrics by judicious choice of materials and surface profiles.
In conventional lens design, the aperture size of an entrance lens or optical stop (i.e., the primary aperture) often determines the effective aperture size of all subsequent lens surfaces (i.e., the secondary aperture) in the lens system. The use of multiple lenses and apertures enables a lens system to simultaneously create an effective focal length and magnification appropriate to the imaging task at hand, reduce image aberrations, and provide correct image orientation. Secondary apertures are typically matched to the effective cross section of the magnified or demagnified entrance aperture propagated through the lens system. In systems with low aberration, the size of the entrance aperture often determines angular resolution of the lens system while the size of the secondary apertures determines the field-of-view of the lens system.
Simple cameras typically balance field-of-view and resolution by using a sequence of lenses having approximately equally sized apertures. Microscopes, on the other hand, achieve large field-of-view and high angular resolution by increasing secondary aperture relative to the collection aperture. Telescopes achieve extra-ordinary angular resolution with a limited field-of-view by decreasing secondary aperture size. Wide-field cameras achieve large field-of-view by tolerating significant aberration across the image with approximately equal primary and secondary apertures. Conventional lens design, therefore, normally requires trade-offs between desired performance metrics. For example, telescopes achieve high angular resolution by sacrificing field-of-view, wide-field imagers achieve large angular fields-of-view by sacrificing diffraction-limited angular resolution, and compound-optics cameras achieve high quality by expanding system volume to include more aberration-correction optics.
In order to overcome some of the limitations of standard imaging optics, multi-aperture cameras have been developed. In multi-aperture systems, a standard cameran objective lens is replaced by an array of lenslets, wherein each lenslet has a reduced focal length in comparison to a conventional camera. In such approaches, a detector measures a set of sub-sampled versions of the object within the field-of-view. Post-processing algorithms are used to generate a high-resolution image from the set of sub-sampled sub-images. The result is reduced system volume; however, the reduction in system volume is achieved at the cost of significant computational post-processing and compromised image quality.
In addition, the design space for multi-aperture cameras is severely restricted, which has limited their adoption in practical systems. The use of a multi-aperture camera requires that the size of its detector array and system aperture be approximately the same size. As a result, conventional multi-aperture designs are generally restricted to very small collection apertures. This also limits the number of camera formats that can be designed. Further, a multi-aperture camera typically has a restricted field-of-view due to a need to prevent the overlapping of sub-images on the detector array. Such overlapping can be avoided by introducing a field stop in the optical design; however, this increases system volume. Alternatively, absorbing barriers can be placed between the sub-image regions of the detector array; however, this significantly increases manufacturing cost and complexity.
For these reasons, a lens system that avoids some of the design trade-offs associated with conventional lens design and that achieves high performance cost-effectively is desirable.